FAULT DIAGNOSIS OF RADIO FREQUENCY LOW NOISE AMPLIFIER USING GMM-HMM

Radio Frequency (RF) analog circuit faults often occur in broadband, high voltage and high temperature environment, so the detection of faults location and prediction of faults occurring time are important research subject.As GMM-HMM (Gaussian Mixture-Hidden Markov Model) algorithm has strong capability for sequential data processing, it is proposed in this paper a fault diagnosis method based on GMM-HMM and has been successfully applied to the fault diagnosis of low noise amplifier circuit based on ATF54143 transistor. Through multiple testing datasets, the experimental results have verified the feasibility of proposed method, the fault model has 100% accuracy for open circuit and short circuit fault diagnosis of the radio frequency circuit.